Termsslug: deliberative-alignment

Deliberative Alignment

审议式对齐

An alignment strategy where models explicitly reason over safety specifications to determine outputs.

Curated Intelligence

1 items · Page 1/1
OpenAI 官方博客T1·Models
85 pts
OpenAI Introduces Deliberative Alignment for Safer Reasoning Models

OpenAI introduces 'Deliberative Alignment' for its o-series models, a new strategy that teaches models to directly learn safety specifications and reason over them. By combining supervised fine-tuning with reinforcement learning, this approach enables models to explicitly weigh safety rules before generating responses, significantly reducing harmful outputs.